Dietary Practices and Associated Factors Among Pregnant Women in West Gojjam Zone, Northwest Ethiopia

Home , Achefer, Bahir Dar Zuria, Gojjam, West Gojjam Zone

Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 https://doi.org/10.1186/s12884-019-2702-z

RESEARCH ARTICLE Open Access Dietary practices and associated factors among pregnant women in West Gojjam Zone, Northwest Ethiopia Yeshalem Mulugeta Demilew1*, Getu Degu Alene1 and Tefera Belachew2

Abstract Background: The optimal dietary practice is a critical requisite for maternal nutrition. However, the majority of Ethiopian pregnant women have inadequate nutrient intakes. These may be due to their poor dietary habits. Identifying factors affecting the dietary practices of pregnant women is crucial to design appropriate interventions. In this country, the dietary practices of pregnant women and determinants are not well studied. Therefore, the purpose of this study was to assess the dietary practices and associated factors among pregnant women in West Gojjam Zone, Northwest Ethiopia. Methods: A community-based cross-sectional study was carried out among 712 pregnant women from May to August 2018. Quantitative data complemented with a qualitative method. Pregnant women were selected using a cluster sampling technique. Structured questionnaires were utilized for data collection. Data were entered into Epi- Info version 7.2.2 and exported to SPSS version 23 software for analysis. Data were described using frequencies and mean. A logistic regression analysis was done. Three focus group discussions and 17 key-informant interviews were conducted for the qualitative data. Focus group discussion participants were mothers, husbands, and health professionals. Typical case and homogeneous sampling techniques were used for the key-informant interviews and focus group discussions, respectively. Thematic analysis was used for the qualitative data. Results: Only 19.9% of respondents had appropriate dietary practices. On the multivariable logistic regression analyses, being food secure [AOR = 2.25, 95% CI: (1.1, 4.5)], having high edible crop production [AOR = 2.00, 95% CI: (1.2, 3.2)] and a favorable attitude [AOR = 1.69, 95% CI: (1.1, 2.6)] were significantly associated with the appropriate dietary practices of pregnant women. In the qualitative study, lack of knowledge on maternal diet, cultural prohibition, and knowledge gap of the professionals were barriers that interfere with dietary practices during pregnancy. Conclusion: Pregnant women in the study area are found to have suboptimal dietary practices. Therefore, health professionals should give regular nutrition counseling using cards and role models for promoting diversified food production and consumption. Keywords: Pregnant women, Dietary practice, Dietary diversity, Food variety score

* Correspondence: [email protected] 1School of Public Health, College of Medicine and Health Sciences, Bahir Dar University, P.O. Box 79, Bahir Dar, Ethiopia Full list of author information is available at the end of the article

© The Author(s). 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 2 of 11

Introduction Among these factors, individual (psychological and During pregnancy, women’s bodies undergo anatomical, socio-demographic) and environmental (economic, so- physiological, and biochemical changes [1, 2]. These bio- cial, cultural, availability, and accessibility of nutritious logical changes increase women’s nutrient requirements food) factors are amendable [21, 22]. Socio-demographic [2]. Thus, pregnant women should eat diversified food- factors were assessed by previous researchers, whereas stuffs that contain an adequate amount of energy, pro- psychological and environmental factors affecting dietary tein, vitamins, minerals, and water [3]. practices of pregnant women were not assessed in the Despite this, the majority of pregnant women in devel- Ethiopian context. oping countries have inadequate nutrient intakes com- Thus, this study adds to the literature on predictors of pared to the standard recommended by the World dietary practice by assessing environmental (enabling) and Health Organization (WHO) [4, 5]. Especially, diets of psychological (predisposing and reinforcing) factors. Enab- Asian and African pregnant women are predominantly ling factors were livestock possession, edible crop, vege- cereal-based with infrequent consumption of animal table, and cash crop production. Predisposing factors were products, vegetables, and fruits [6]. perceived susceptibility to and severity of the conse- Few studies done in Ethiopia have shown that nutrient quences of poor maternal diet, attitude, and intention to intakes of pregnant women aren’t sufficient to meet their take a balanced diet, behavioral control, perceived benefit, nutrient demand. Nutrient inadequacies are as a result and barrier of taking adequate diet during pregnancy. Re- of poor dietary practices (observable action of dietary inforcing factors were family support to take a healthy habits) [7, 8]. According to the literature, good dietary diet, women’s autonomy, and subjective norms. practices of pregnant women vary within the country About 34.5% of pregnant women who attend antenatal ranging from 26.9% in the Ambo district to 40.1% in the care at public hospitals of Addis Ababa, Ethiopia Gondar town [9–12]. reported good dietary practices. The same study also To assess the dietary practices of pregnant women, identified education, monthly income, attitude, and gra- some of the literature in the country used frequency of vidity as factors associated with the dietary practices of meal, while others utilized questions on the nutrient con- pregnant women. This study used the above study as a tents of the food which is not feasible to use in the Ethiop- reference to calculate the sample size using single and ian context since the majority of pregnant women didn’t double population proportion formulas. know the nutrient contents of the food. Nutrition infor- Identifying factors that affect the dietary practices of mation, education, monthly income, dietary knowledge, pregnant women is timely to design appropriate inter- family size, age, ownership of radio, having an illness, and ventions. Therefore, the aim of this study was to assess husband occupation were independent factors for the the dietary practices and associated factors among preg- dietary practices of pregnant women [9, 10, 12, 13]. nant women in West Gojjam Zone. The result will be a Inadequate dietary intakes are associated with intrauter- helpful baseline data about pregnancy & nutrition in the ine growth retardation, low birth weight, and premature Amhara region, as it may be the first study of its kind in delivery [14, 15]. These problems remained the foremost the region. It could be used as an input to policymakers public health concern in Ethiopia and associated with the and planners at the national, regional, and zonal level to high neonatal and infant mortalities [16]. Conversely, ex- design nutrition interventions important for improving cessive consumption of energy-dense foods linked with maternal diet during pregnancy. The health service pro- excessive gestational weight gain [17, 18] that, in turn, in- viders can take account of the findings of this study to creases the risk of adverse pregnancy outcomes [19]. modify the way nutrition education is delivered. Taking this into consideration, the Ethiopian government continuously showed its commitment to avert nu- Methods tritional problems by developing food and nutrition Study area and period policy, strategy, and programs. Furthermore, the govern- The study was conducted in West Gojjam Zone from ment designed the “Seqota Declaration” to end stunting May to August 2018. The zone is one of the 11 zones in by 2030 [20]. the Amhara Region. In Ethiopia, region and zone are the Despite the effort of the government to improve nutri- first and second level administrative divisions, respect- tion, few studies done on maternal nutrition in the ively. Zones are further divided into a number of wore- country revealed suboptimal nutrient intakes of pregnant das. Woredas (the third-level administrative divisions in women compared with the standard recommended by the country) further divided into kebeles (the smallest the WHO [7, 8]. administrative unit in the country). The dietary practices of pregnant women are influ- West Gojjam Zone has 15 woredas with a total popu- enced by a set of complex biological, physiological, psy- lation of 2,641,240, of which 50.7% were females. The chological, economic, social, and environmental factors. number of estimated pregnant women was 61,072. Teff, Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 3 of 11

maize, millet, bean, pea, grass pea, pepper, barley, wheat, eight woredas were excluded from the study because cabbage, collard green, tomato, potato, papaya, mango, dietary practices of pregnant women in these woredas and avocado are foodstuffs commonly produced in the may not be representative of the dietary practices of all zone. Cattle, chicken, sheep, and goat are the common pregnant women. livestock in the study area [23]. From the seven woredas, three (Bahir Dar Zuria, South Achefer, and Burie Zuria) woredas were selected by a Study design simple random sampling (SRS) method. Based on pro- A community-based cross-sectional study was done portional to the size allocation, ten clusters from Bahir among pregnant women. In this study, quantitative data Dar Zuria woreda, six clusters from South Achefer wor- complemented with a qualitative method. A quantitative eda, and another six clusters from Burie Zuria woreda data collection followed by a qualitative method to ex- were also selected using the SRS technique. plore and explain the result of the quantitative data. House to house survey was conducted to identify eligible pregnant women in selected clusters. All eligible Source population and study population women were included in the study. The women’s preg- All pregnant women in the zone were the source popu- nancy status was assessed by inquiring about her last lation, whereas pregnant women in the selected kebeles menstrual period and confirming with a pregnancy test. were the study population. Data collection and measurements Inclusion criteria Data were collected by interviewer-administered question- Since dietary practice is affected by the local culture, preg- naires. The questionnaire includes socio-demographic nant women who lived in the kebele at least for 6 months variables, obstetrics history, dietary related variables, food se- were included in this study. Gestational age before 16 curity status, maternal perception, and socio-cultural issues. weeks of gestation was another inclusion criterion because Six female nurses and three male public health professionals this is a baseline data of a cluster-randomized controlled were recruited as data collectors and supervisors, respect- community trial, which was registered in Clinical Trials.- ively. Additionally, three female laboratory technicians were gov with the identification number of NCT03627156. Title recruited to do pregnancy tests. Data collectors administered of the trial is “the effect of guided counseling in improving thequestionnairethroughaface-to-face interviews at the selected nutrition outcomes of pregnant women in West participants' homes. To maintain the optimal privacy of the Gojjam Zone, Ethiopia: A cluster randomized controlled mothers, other family members didn’thavefreeaccesstothe community trial.” place where the interviews were conducted. Food frequency questionnaire (FFQ), which was taken Exclusion criteria from previously validated questionnaires [24–26], con- Pregnant women who had confirmed or diagnosed taining 54 food items was used to collect dietary data. hypertension and/or diabetes mellitus were excluded The questionnaire was also validated after the assess- from the study because nutrient requirements and diet- ment of locally available foodstuffs. For the food fre- ary practices of hypertensive and diabetes cases are dif- quency questions, the women were asked about the ferent from their counterparts. frequency of consumption of each food per day, per week or per month in the prior 3 months by taking the Sample size determination variation of dietary intakes within days of the week into The sample size was calculated using single and double consideration [25, 26]. population proportion formulas with their respective as- Food items of the FFQs were grouped into nine food sumptions. The sample size calculated by a single popula- groups: 1. cereals, roots, and tubers; 2. Vitamin-A-rich tion proportion formula using the following assumptions fruits and vegetables; 3. other fruits; 4. other vegetables; gave the largest sample size used in this study: 95% confi- 5. legumes and nuts; 6. meat, poultry, and fish; 7. fats dence level, 34.5% proportion of appropriate dietary prac- and oils; 8.dairy; and 9. Eggs [24].The consumers of a tice based on previous study [13], 5% marginal error, 5% food item were defined as the consumption of a food non-response rate, and design effect 2. The calculated item at least once a week [26]. The number of food sample size was 712. groups the women ate within a week were counted to analyze dietary diversity score (DDS). Sampling procedure Food variety score (FVS) was computed by counting A cluster sampling technique was used to select preg- the individual food items the women consumed within a nant women. From the 15 woredas in the zone, eight of week. Then, the mean FVS was analyzed. The utilization them had a nutrition intervention program implemented of animal source food (ASF) was assessed by counting by a nongovernmental organization. Therefore, these the frequency of each animal source foods the women Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 4 of 11

took within the days of a week. Finally, the frequency of tercile was labeled as a favorable attitude, if not unfavor- ASF consumption was divided into terciles (three parts). able attitude. Dietary diversity score, FVS, ASF consumption, and fre- Subjective norms, intention, perceived susceptibility, quency of meal were used to assess dietary practices. perceived severity, perceived benefit, and perceived bar- The food security status of the household was assessed riers were assessed using their respective composite using 27 questions, which were adapted from the house- questions. Mean was computed and women who scored hold food insecurity access scale. The questions were pre- above the mean for each variable were categorized as viously validated for use in developing countries [27]. having subjective norms, intention, perceived suscepti- Food secure households experienced fewer than the first 2 bility, perceived severity, perceived benefit, and per- food insecurity indicators. Whereas, a household which ceived barriers, otherwise no. experienced from 2 to 10, 11–17, and > 17 food insecurity indicators were considered as mildly, moderately, and se- Quality assurance mechanisms verely food insecure households, respectively. To maintain the quality of the research, the question- The wealth index of the household was determined naire was adapted from standard data collection instru- using Principal Component Analysis (PCA) by consider- ments, and it was pretested. Data collectors, supervisors, ing latrine, water source, household assets, livestock, and and laboratory technicians were trained for 3 days. The agricultural land ownership. The responses of all non- data collection procedure was supervised by the supervi- dummy variables were classified into three parts. The sors and principal investigator. highest score was coded as 1. Whereas, the two lower The data collection team held a daily meeting, and values were given code 0. In PCA, those variables having feedback was given daily. Double entry and verification a commonality value of greater than 0.5 were used to were done by the principal investigator. Important produce factor scores. Lastly, the score for each house- assumptions were checked using the standard proce- hold on the first principal component was retained to dures. The dietary practice was the dependent variable create the wealth score. Quintiles of the wealth score for the quantitative data. Socio-demographic, enabling, were created to categorize households as poorest, poor, predisposing, reinforcing, and obstetric factors were in- medium, rich, and richest. dependent variables. To determine edible crop and vegetable production, each crop and vegetable species cultivated by the house- Data processing and analysis hold were counted. The number of crops and vegetables Quantitative data were edited and coded manually. Data produced was classified into three parts. The highest were entered into Epi Info version 7.2.2 software and value was labeled as high production, while the two exported to SPSS version 23 software for cleaning and lower values were considered as low production. In the analysis. The frequency of meal, FVS, ASF consumption, study area, khat, eucalyptus, “Gesho” (a local plant used DDS, and dietary practice were determined. to make tella or local beer and areki), pepper, and onion Since a cluster sampling technique was used to select are the common cash crops. Count of these cash crops the study participants, a generalized linear mixed model was used to decide cash crop production. was fitted to include cluster-level variables. The intercept- The total ownership of livestock was measured by only model estimates the intercept as 0.272 (the average Tropical Livestock Units (TLUs). There is no TLU appropriate dietary practice across all clusters was 0.272 index created specifically for Ethiopia, therefore, the in- but wasn’t statistically significant (p =0.068)).Theintra- dexesforTropicalAfricawasusedinthisstudy.The cluster correlation coefficient was closer to zero (0.077). TLUs were calculated using the following weighted This showed that 92.3% of the dietary practice scores were index factors: cattle = 0.7, horses = 0.5, mules = 0.5, don- explained by women level variables. The non-significant key = 0.5, sheep = 0.1, goats = 0.1, chickens = 0.01 [28]. variability in dietary practices of pregnant women at the Women’s autonomy was assessed using eight ques- cluster level could be due to considering the design effect tions. For each question, code one was given when a de- of 2 during sample size calculation, which in turn in- cision was made by the woman alone or jointly with her creased the calculated sample size. Therefore, bivariate husband, otherwise zero. The mean was used to classify and multivariable logistic regression analyses were used to a woman’s decision making power [16]. assess predictors. Hosmer-Lemeshow Goodness-of-fit-test Maternal knowledge on diet during pregnancy was was done to check model fitness. Correlation between in- assessed using 12 questions. Code one was given for dependent variables was checked using the Pearson Cor- each question when the response was correct, or else relation Coefficient. P-value < 0.2 was used as a cut-off zero. The attitude was assessed by 20 Likert scale ques- point to select variables for the final model. Backward tions using PCA. The factor scores were summed and elimination was used, and P-value < 0.05 was considered ranked into terciles (three parts). Then the highest statistically significant. Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 5 of 11

Definitions of terms The qualitative data were collected using the semi- structured interview and FGD guides, prepared by ➢ Food is any nutritious substance that people eat reviewing existing literature and the local context. More- or drink over, the guides were revised based on identified gaps ➢ Appropriate dietary practice: when women had at during interviewing or facilitating. During data collec- least four meals daily, good FVS, high DDS, and high tion, tape-recorder was used in addition to field notes. ASF consumption, whereas it was inappropriate when All interviews were conducted by the principal investiga- women had less than four meals daily or poor FVS or tor at a convenient time in public meeting halls in the low DDS or low ASF consumption [25]. village. The interviews took from 30 to 45 min. The ➢ High DDS: tertiles were calculated from food FGDs were also facilitated by the principal investigator groups, and the highest tertile was considered as a high and took from 45 min to 1 h. Notetaker who participated DDS, whereas the rest two lower tertiles were taken as in qualitative data collection before this research was re- a low DDS [26]. cruited to take notes during the interviews and FGDs. ➢ Food variety score: women who had above the Transcription was done on the same day of data col- mean food variety were considered as having good FVS, lection. During the interpretation of findings, individuals’ otherwise having poor FVS [25]. detailed views of the meaning of events were considered. ➢ Utilization of ASF: the highest tertile for ASF To assure trustworthiness, key-informant interviews consumption was considered as the high frequency of were triangulated with FGDs. The quantitative data were ASF consumption, whereas the two lower tertiles were triangulated with the qualitative method. taken as the low frequency of ASF consumption [26]. First, the researchers repeatedly listened to tape- ➢ Knowledge of the women about maternal diet: recorded interviews and FGDs. Then, verbatim tran- women who score > =75 and > 50% from knowledge scription was done in the Amharic language. Tran- questions were considered as having high and medium scribed interviews and FGDs were translated from knowledge scores, respectively. Or else labeled as Amharic to English. Finally, data were entered into the having low knowledge score [12]. QDA MINER LITE version 2.0.2 software for analysis. Codes were given for similar ideas using the software and cross-checked between researchers, used to establish Qualitative data analytical categories. Then, the meanings were organized The dietary practice is an observable action of dietary into themes for thematic analysis. Finally, the result was habit, which is the lived experience of an individual. presented in narratives in triangulation with the quanti- Therefore, a phenomenological qualitative study was con- tative results and illustrations. ducted using key-informant interviews and focus group discussions (FGDs). Three FGDs involving 6–12 partici- Ethical consideration pants in each group were conducted in Amharic (local) Institutional Review Board of Bahir Dar University ap- language. Twelve FGD participants were mothers. Eight proved the study. Permission was obtained from West FGD participants were husbands, and six were health Gojjam Zone and each woreda administrators. Written professionals. consent (fingerprint for women who couldn’t read and Seventeen key-informant interviews were carried out write) was secured from the study participants. Confi- among pregnant women, health professionals, nutrition dentiality was maintained by excluding personal identi- officers (nurses and pharmacist in their profession), fiers from the data collection form and keeping the data community leaders and one-to-five network member in a locked board. leaders (one-to-five network is a lower level governmen- tal structure consisting of five people. One leader moni- tor, mobilize and train five members about health, Results nutrition, and developmental related issues). All partici- Socio-demographic and obstetric characteristics of pants were Orthodox Christians and Amhara in Ethni- pregnant women city. The respondents’ age ranged from 25 to 55 years. Seven hundred twelve eligible pregnant women were The educational status of the qualitative study partici- invited to participate in the study. Complete data were ob- pants ranged from having no formal education to a pri- tained from 694 study participants, with a response rate of mary degree (Table 1). 97.5%. The mean (± SD) age of women was 28.14 (±5.97) A typical case and extreme sampling techniques were years. All respondents were Amhara in ethnicity and rural used to select the study participants for the key- dwellers. Almost all (99.4%) women were Orthodox Chris- informant interviews. FGD participants were chosen tian by religion. The wealth index showed that only 19.6% using a homogeneous sampling technique. of the respondents were in the richest category. One in Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 6 of 11

Table 1 Socio-demographic characteristics of qualitative study In this study women who had medium level of know- participants in West Gojjam Zone, Ethiopia ledge were 1.81 times [AOR = 1.81, 95% CI: (1.1, 3.1)] Variable Key-informant Focus-group-discussions and those who had high level of knowledge were 3.23 interviews (n = 17) (n = 26) times [AOR = 3.23, 95% CI: (1.9, 5.5)] more likely to Age in years have appropriate dietary practices compared with those <=30 7 5 who had low level of knowledge. The likelihood of hav- 31–34 5 7 ing appropriate dietary practice was 2.25 times higher 35–39 1 8 among food secure women [AOR = 2.25, 95% CI: (1.1, 4.5)] than their counterparts. >=40 4 6 Women who reside in the household with high edible Sex crop production were 2.00 times [AOR = 2.00, 95% CI: Male 8 14 (1.2, 3.2)] and those who had high vegetable production Female 9 12 were 1.50 times [AOR = 1.50, 95% CI: (1.1, 2.2)] more Occupation likely to have appropriate dietary practices compared Farmer 5 20 with their counterparts. Women with a favorable attitude had 1.69 times higher odds of having appropriate Clinical Nurse 4 dietary practice than women with an unfavorable atti- Midwife 5 6 tude [AOR = 1.69, 95% CI: (1.1, 2.6)]. In this study edu- Pharmacist(officer) 1 cational status, wealth index, livestock possession, and Doctor 2 cash crop production had no association with the dietary Educational status practices of pregnant women (Table 4). ,t No formal education 6 15 In the qualitative findings he three identified central themes were dietary practice, barriers of having appro- Primary education 1 5 priate dietary practice, and suggested solutions to im- Diploma 7 6 prove maternal diet. Degree 3 All respondents in the key-informant interviews and Marital status FGDs reported poor dietary practices of pregnant Married 4 6 women. A female key-informant explained: pregnant Unmarried 13 20 women frequently eat millet and maize injera (fermented pancake-like flatbread) with shero (stew). The amount and frequency of meals are less during pregnancy compared with their meals before pregnancy… five (20.5%) women had first pregnancy, and the mean (± A one-to-five-network leader, female key-informant SD) parity of women was 2.49 (±1.96) (Table 2). also said: though all foodstuffs are available at home, pregnant women eat maize injera with legume stew… Dietary practices of pregnant women Barriers that hamper the dietary practices of pregnant The dietary practices of pregnant women were shown in women Table 3. The most frequently eaten foods were cereals, The study participants mentioned a lack of knowledge, legumes, and oils but animal source foods, fruits, and food insecurity, cultural prohibition, workload, unfavor- vegetables were not often consumed by the study partici- able attitude, health professionals’ knowledge gap, and pants (Fig. 1). negligence as barriers affecting the dietary practices of pregnant women. Factors associated with dietary practices of pregnant Thirteen key-informants and all FGD participants re- women ported a lack of knowledge on diet during pregnancy as In the logistic regression analysis, Hosmer-Lemeshow a predominant obstacle for having good dietary practice. Goodness-of-fit-test was 0.812. Pearson Correlation Co- A midwife key-informant reported: in this area, pregnant efficient was 0–0.2. Table 4 showed factors that had a women have a low level of knowledge on maternal diet. statistically significant association with the dietary prac- Due to this, they sell their nutritious foods and eat tices of pregnant women. Multivariable logistic regres- roasted grains throughout their pregnancy. sion analysis revealed food security, knowledge, attitude, The majority of the qualitative study participants edible crop, and vegetable production as the factors that stated that food insecurity was the main constraint of were significantly associated with the dietary practices of having appropriate dietary practice. A pregnant woman pregnant women. key-informant stated: though this area is food secure, Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 7 of 11

Table 2 Socio-demographic and obstetric characteristics of Table 2 Socio-demographic and obstetric characteristics of pregnant women in West Gojjam Zone pregnant women in West Gojjam Zone (Continued) Variables Frequency Percentage (%) Variables Frequency Percentage (%) (n = 694) (n = 694) Age (years) Parity < 20 42 6.1 0 153 22.0 20–24 138 19.9 1–3 331 47.7 25–29 207 29.8 4–5 159 22.9 30–34 171 24.6 > =6 51 7.4 > =35 136 19.6 Mean (+SD) 28.14(+ 5.97) there are some women with poor living standards, for Religion these women, food insecurity is the bottleneck to have ap- Orthodox 690 99.4 propriate dietary practices. Muslim 4 0.6 Six key-informants and almost all FGD participants described the negative consequences of a cultural pro- Educational status hibition on maternal diet. There is no special diet cultur- Cannot read and write 520 74.9 ally recommended for pregnant women in the study area Can read and write 42 6.1 rather the community suggested reducing the amount of Primary education 84 12.1 meal per serving and avoiding some selected foodstuffs Secondary education 48 6.9 (linseed, pumpkin, and chickpea) during pregnancy. Occupational status Moreover, taking an additional meal during pregnancy compared with the frequency of meals in her non- Housewife 355 51.2 pregnant state was discouraged by the community. This Farmer 339 48.8 cultural prohibition affects the dietary practices of Marital status pregnant women. The majority of FGD participants re- Married 687 99.0 ported: the community suggests reducing consumption of Single/ divorced 7 1.0 food by considering that it predisposes to a big baby. A Husband education (n = 687) pregnant woman key-informant said that pregnant women avoid linseed, pumpkin, and chickpea due to cul- Cannot read and write 344 50.1 tural restriction... Can read and write 176 25.6 Primary education 120 17.5 Secondary and above education 47 6.8 Table 3 Dietary practice of pregnant women in West Gojjam Husband occupation (n = 687) Zone, Ethiopia Farmer 686 99.9 Variables Frequency (n = 694) Percentage (%) Government employee 1 0.1 Dietary practice Wealth index Appropriate 138 19.9 Poorest 126 18.1 Inappropriate 556 80.1 Poor 151 21.8 Women dietary diversity Medium 139 20.0 High 229 33.0 Rich 142 20.5 Low 465 67.0 Richest 136 19.6 Food variety score Family Size Good (above the mean) 322 46.4 < 5 358 51.6 Poor (mean & lower) 372 53.6 > =5 336 48.4 Utilization of animal source food Gravidity High 254 36.6 1 142 20.5 Low 440 63.4 2–3 199 28.7 Frequency of meal/day 4–5 218 31.4 > =4 245 35.3 > =6 135 19.4 < 4 449 64.7 Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 8 of 11

Fig. 1 Consumption of foods by pregnant women based on food groups in West Gojjam Zone, Ethiopia

Another reason for having inappropriate dietary practice Possible solutions to improve maternal diet during during pregnancy was the habit of practicing fasting in pregnancy Orthodox Christians. Nearly all FGD participants reported: All respondents recommended the need for improving in this area, pregnant women abstain from eating animal counseling methods. They proposed trimester based products, and they do not take breakfast during fasting days. counseling. Qualitative study participants recommended These inhibit them from taking an adequate diet. using role models, counseling cards, and guidelines dur- According to the qualitative study findings, a low level ing counseling of pregnant women. Study participants of crop production decrease availability and accessibility also proposed developing nutrition guidelines with detail of foodstuffs in the household. These, in turn, deter- information about diet during pregnancy. Moreover, mined women’s dietary practices. All FGD participants they suggested assigning nutritionists starting from the explained: we can’t produce a variety of crops due to the health center to a regional level. Furthermore, respon- shortage of land and the high cost of fertilizers. This re- dents proposed incorporating nutrition education in a stricts women from taking a variety of foodstuffs. school curriculum and giving refreshment training to Almost all respondents reported excessive workload and health professionals. negative attitude of the women as constraints for having ap- A nutrition focal key-informant suggested: the method propriate dietary practices during pregnancy. The majority of of nutrition education should be modified to improve the FGD participants claimed: mostpregnantwomendofarmac- maternal diet. The counseling should be based on the tri- tivities in addition to their household duties, which hinder mester of pregnancy. It needs to be supported by role them from taking their meals on time. A significant number models, counseling cards, and guidelines. To do so, cap- of FGD participants said: pregnant women are not interested acity building training should be given to health profes- in preparing and eating their own diversified meals. They pre- sionals. Besides, there is a need to assign a nutritionist ferred eating a monotonous diet. with sufficient knowledge about nutrition at each level in All respondents reported a knowledge gap of profes- the health system. Another nutrition focal key-informant sionals and negligence as the foremost obstacles to hav- recommended as the best solution to improve maternal ing appropriate dietary practice during pregnancy. A nutrition is incorporating nutrition education in the nutrition focal key-informant stated: I am not doing any- school curriculum… thing to improve the maternal diet since I have no enough information about maternal nutrition. Not only I but also all other responsible bodies are not giving atten- Discussion tion to maternal nutrition. Inappropriate dietary practice during pregnancy has A female, midwife key-informant said: we give less time been linked to the risk of adverse pregnancy outcomes. for nutrition counseling compared with other packages. Despite this fact, in this study, only 19.9% [95% CI: We advise the women to take what is available at home (16.7, 22.6)] of respondents had appropriate dietary by adding one extra meal because we don’t have enough practices. Similarly, qualitative study participants re- knowledge about maternal nutrition. ported poor dietary practices of pregnant women. Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 9 of 11

Table 4 Factors associated with dietary practices of pregnant Table 4 Factors associated with dietary practices of pregnant women in West Gojjam Zone, Ethiopia women in West Gojjam Zone, Ethiopia (Continued) Variable Dietary practice COR AOR Variable Dietary practice COR AOR (95% CI) (95% CI) (95% CI) (95% CI) Appropriate inappropriate Appropriate inappropriate Get nutrition education Low 66(9.5) 386(55.6) 1.00 1.00 Yes 56(8.1) 140(20.1) 2.02(1.3,2.9) AOR Adjusted odds ratio, COR Crude odds ratio, 95%CI 95 % confidence interval No 82(11.8) 416(60.0) 1.00 Food secure This result is much lower than the study findings in Yes 127(18.3) 417(60.1) 3.84(2.0,7.3) 2.25(1.1,4.5) Northwestern [12] and Eastern Ethiopia [29]. The dis- No 11(1.6) 139(20.0) 1.00 1.00 crepancy might be due to differences in the study set- Intention tings since the current study was conducted among rural Yes 95(13.7) 260(37.5) 2.51(1.7,3.7) residents with little access to nutrition education and health services. Moreover, respondents in this study had No 43(6.2) 296(42.6) 1.00 a low level of education. Their literacy level might have Knowledge on diet been a barrier to access information related to good nu- High 65(9.4) 108(15.6) 5.61(3.4,9.1) 3.23(1.9,5.5) trition practices. Medium 42(6.0) 159(22.9) 2.46(1.5,4.1) 1.81(1.1,3.1) Only 33.0% [95% CI: (29.5, 36.7)] of the study partici- Low 31(4.5) 289(41.6) 1.00 1.00 pants consumed a diversified meal. The minimum FVS Attitude was 46.4% [95% CI: (42.4, 50.3)]. Moreover, 36.6% [95% CI: (33.0, 40.0)] of respondents consumed ASFs. These Favorable 76(11.0) 154(22.2) 3.20(2.1,4.6) 1.69(1.1,2.6) findings are in line with previous study findings in devel- Unfavorable 62(8.9) 402(57.9) 1.00 1.00 oping countries [30]. According to the qualitative study Wealth index findings, pregnant women frequently took a cereal-based Richest 42(6.1) 94(13.5) 3.30(1.7,6.3) monotonous diet. Moreover, they occasionally took Rich 34(4.9) 108(15.6) 2.33(1.2,4.5) fruits, vegetables, and animal products. These findings Medium 21(3.0) 118(17.0) 1.31(0.6,2.6) correlated with the study done in Northeast Ethiopia [31]. This indicates poor quality of maternal diets, which Poor 26(3.7) 125(18.0) 1.53(0.7,3.0) may significantly affect pregnancy outcomes [32]. Poorest 15(2.2) 111(16.0) 1.00 Knowledge about diet during pregnancy plays a central Family support role in determining women’s dietary practices [33, 34]. Yes 66(9.5) 189(27.2) 1.78(1.2,2.5) In this study, the likelihood of having appropriate dietary No 72(10.4) 367(52.9) 1.00 practice increased with increased maternal knowledge Edible crop production about diet during pregnancy. All respondents in the qualitative study claimed a lack of knowledge as a barrier High 52(7.5) 91(13.1) 3.13 (2.0,4.7) 2.00(1.2,3.2) that hampers the dietary practices of pregnant women. Low 86(12.4) 465(67.0) 1.00 1.00 This finding is in agreement with the study findings in Household decision making Bahir Dar Town [12] and Addis Ababa [13]. However, a Yes 105(15.1) 358(51.6) 1.76(1.1,2.7) study finding in Cameron reported a lack of association No 33(4.8) 198(28.5) 1.00 between knowledge and dietary practice, which is differ- Perceived susceptibility ent from our study finding [35]. Production of edible crops and vegetables had an asso- Yes 92(13.3) 263(37.9) 2.22(1.5,3.2) ciation with a better maternal diet. Limited production No 46(6.6) 293(42.2) 1.00 of crops and vegetables were barriers to having an ap- Perceived severity propriate maternal diet in the qualitative study findings. Yes 112(16.1) 374(53.9) 2.09(1.3,3.3) This finding is in line with the study finding in Southern No 26(3.8) 182(26.2) 1.00 Ethiopia [36]. The possible explanation might be that Perceived benefit the crop and vegetable production affect the availability and accessibility of foodstuffs, which is the direct con- Yes 112(16.1) 345(49.7) 2.63(1.6,4.1) tributor for a good maternal diet [7, 22]. No 26(3.8) 211(30.4) 1.00 The likelihood of having appropriate dietary practices Vegetable production was higher among women who have a favorable attitude High 72(10.4) 170(24.5) 2.47(1.7,3.6) 1.50(1.1,2.2) compared with their counterparts. This finding is supported by the qualitative study finding in which the Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 10 of 11

negative attitude of the women was the reason for hav- Recommendations ing a poor diet during pregnancy. This might be due to Nutrition education intervention needs to be tailored to the fact that women who have a favorable attitude to- meet the need for pregnant women to improve their wards maternal diet are more likely to have the intention dietary practices. Therefore, health professionals should to take a balanced diet that directly affects their dietary give regular nutrition counseling using cards and role practices [37, 38]. models. Nutritionists should be assigned at each level in Food secured women had 2.25 times higher probability the health system. Incorporating nutrition education into of eating adequate quantity and a high-quality diet com- the school curriculum is recommended. Crop and vege- pared with their counterparts. Similarly, in the qualitative table production using available resources and backyard study, the predominant reasons for poor maternal diet gardening should be encouraged by liaising with the were unavailability and inaccessibility of foodstuffs. As per agricultural sector. There is also a need to refresh the our study and previous study findings [39], food security knowledge of health professionals on diet during is the primary prerequisite to improve maternal diet [40]. pregnancy. There was no association between livestock possession and dietary practices of pregnant women. The reason for Abbreviations ASF: Animal source food; DDS: Dietary diversity score; FFQ: Food frequency this might be a high cost of animal products compared questionnaire; FGDs: Focus group discussions; FVS: Food variety score; with plant foodstuffs. In this study, cash crop production GWG: Gestational weight gain; IUGR: Intrauterine growth retardation; was not a predictor for the dietary practices of pregnant LBW: Low birth weight; PCA: Principal Component Analysis; QDA: Qualitative Data Analysis; TLUs: Tropical Livestock Units; WHO: World Health women. This finding is consistent with the study finding Organization in Ghana [41]. This might be due to the reason that cash crops are primarily produced for sale, not for home Acknowledgements consumption. We would like to express our heartfelt gratitude to the study participants since without them this study may not be visible. We acknowledge data This study makes important contributions to under- collectors and supervisors for their diligent contribution. Once again, we stand dietary practices and determinants of diet during would like to thank Dr. Guadie Sharew for his significant role in language pregnancy. It can be used as an input for policymakers, edition and Bahir Dar University for its financial support. planners, researchers, programmers, and health profes- Authors’ contributions sionals in the improvement of maternal diet to achieve YMD: conceived and designed the study, conducted statistical analysis and the “Seqota Declaration” to end stunting by 2030. result interpretation, prepared manuscript. The author read and approved The use of validated FFQ and supplementing the the manuscript. GDA: conceived and designed the study, conducted statistical analysis and result interpretation, prepared manuscript. The author quantitative data with qualitative study findings were the read and approved the manuscript. TB: conceived and designed the study, strengths of this study. Although utmost efforts were conducted statistical analysis and result interpretation, prepared manuscript. made to reduce bias by giving intensive training for data The author read and approved the manuscript. collectors on how to probe the women to remember Funding their dietary practices, the possibility of social desirability This research was funded by Bahir Dar University. The University paid bias may be the limitations of this study. Key-informant perdiem for data collectors, supervisors and laboratory technicians who interviews and FGD facilitations were held by the princi- confirmed pregnancy test but didn’t participate in design, analysis and interpretation of findings. pal investigator with previous experience of qualitative data collection. The make-up of the focus groups may Availability of data and materials have influenced the discussion. But, we believe giving a The datasets analyzed during the current study are not publicly available clear introduction about the objective and probing dur- because this study is a baseline study of a cluster-randomized controlled community trial that was registered in Clinical Trials.gov with the identifica- ing discussion promotes the active participation of all tion number of NCT03627156. The authors have using the baseline data for FGD participants. analysis of other objectives. But the datasets are available from the corre- sponding author on reasonable request.

Ethics approval and consent to participate Conclusion The study was approved by Institutional Review Board of Bahir Dar University. Pregnant women in the study area had suboptimal diet- Permission was obtained in West Gojjam Zone and each woreda administrators. ary practices. Having knowledge and favorable attitude, Written consent (fingerprint for those who couldnot read and write) was secured from the study participants. Confidentiality was maintained throughout the study by crop and vegetable production, residing in food secured excluding personal identifiers from the data collection form and keeping the data in households were predictors for having appropriate diet- alockedboard. ary practice. In the qualitative study, lack of knowledge, food insecurity, cultural prohibition, unfavorable atti- Consent for publication Not applicable. tude, professionals’ knowledge gap, negligence, and lack of a responsible body were reasons to have poor dietary Competing interests practices. The authors declare that they have no competing interests. Demilew et al. BMC Pregnancy and Childbirth (2020) 20:18 Page 11 of 11

Author details 22. Shamim AA, Mashreky SR, Ferdous T, Tegenfeldt K, Roy S, Rahman AKMF, 1School of Public Health, College of Medicine and Health Sciences, Bahir Dar et al. Pregnant women diet quality and its Sociodemographic determinants University, P.O. Box 79, Bahir Dar, Ethiopia. 2Department of Nutrition and in southwestern Bangladesh. Food Nutr Bull. 2016;37(1):14–26. https://doi. Dietetics, Faculty of Public Health, Jimma University, P.O. Box 378, Jimma, org/10.1177/0379572116632137. Ethiopia. 23. West Gojjam Zone Agriculture Bureau. West Gojjam Zone annual report. Fenote Selam: Woreda Health Bureau; 2017. Received: 19 June 2019 Accepted: 24 December 2019 24. FAO and FHI 360. Minimum dietary diversity for women: a guide for measurement. Rome; 2016. 25. Belachew T, Lindstrom D, Gebremariam A, Hogan D, Lachat C, Huybregts L, et al. Food insecurity, food based coping strategies and suboptimal dietary References practices of adolescents in Jimma zone Southwest Ethiopia. PLoS One. 1. Kanadys WM. Maternal underweight and pregnancy outcome: prospective 2013;8(3):e57643. https://doi.org/10.1371/journal.pone.0057643. cohort study. Arch Perinat Med. 2007;13(3):23–6. 26. Workicho A, Belachew T, Feyissa GT, Wondafrash B, Lachat C, Verstraeten R, 2. Rolfers SR, Pinna K, Whitney E. Understanding normal and clinical nutrition. et al. Household dietary diversity and animal source food consumption in 18th ed. Wadsworth: Yolanda Cossio; 2009. Ethiopia: evidence from the 2011 welfare monitoring survey. BMC Public 3. WHO. WHO recommendations on antenatal care for a positive pregnancy Health. 2016;16:1192. https://doi.org/10.1186/s12889-016-3861-8. experience. Geneva: WHO; 2016. 27. Gebreyesus SH, Lunde T, Mariam DH, Woldehanna T, Lindtjørn B. Is the 4. Gernand AD, Schulze KJ, Stewart CP, WestJr KP, Christian P. Micronutrient adapted Household Food Insecurity Access Scale (HFIAS) developed deficiencies in pregnancy worldwide: health effects and prevention. Nat Rev internationally to measure food insecurity valid in urban and rural Endocrinol. 2016;12(5):274–89. https://doi.org/10.1038/nrendo.2016.37. households of Ethiopia? BMC Nutr. 2015;1:2 http://www.biomedcentral.com/ 5. Lee SE, Talegawkar SA, Merialdi M, Caulfield LE. Dietary intakes of women bmcnutr/content/1//2. during pregnancy in low- and middle-income countries. Public Health Nutr. 28. Pica-Ciamarra U, Baker D, Morgan N, Zezza A, Azzarri C, Ly C, et al. Investing in 2012;16(8):1340–53. the Livestock Sector: Why Good Numbers Matter, A Sourcebook for Decision 6. Sun E, Sameera T, Mario M, Laura E. Whate are African women eating Makers on How to Improve Livestock Data. Washington DC: World Bank; 2014. during pregnancy? FASEB J. 2011;25:592. 29. Kedir H, Berhane Y, Worku A. Magnitude and determinants of malnutrition 7. Zerfu TA, Umeta M, Baye K. Dietary habits, food taboos, and perceptions among pregnant women in eastern Ethiopia: evidence from rural, – towards weight gain during pregnancy in Arsi, rural central Ethiopia: a community-based setting. Matern Child Nutr. 2016;12:51 63. qualitative cross-sectional study. J Health Popul Nutr. 2016;35:22. https://doi. 30. Cornelsen L, Alarcon P, Häsler B, Amendah DD, Ferguson E, Fèvre EM, et al. org/10.1186/s41043-016-0059-8. Cross-sectional study of drivers of animalsource food consumption in low- 8. Asayehu TT, Lachat C, Henauw SD, Gebreyesus SH. Dietary behaviour, food income urban areas of Nairobi, Kenya. BMC Nutr. 2016;2:70. https://doi.org/ and nutrient intake of women do not change during pregnancy in 10.1186/s40795-016-0109-z. Southern Ethiopia. Matern Child Nutr. 2016;13(2):e12343. https://doi.org/10. 31. Aliwo S, Fentie M, Awoke T, Gizaw Z. Dietary diversity practice and 1111/mcn. Epub 2016 July 3. associated factors among pregnant women in north East Ethiopia. BMC Res 9. Tolera B, Mideksa S, Dida N. Assessment of dietary practice and associated Notes. 2019;12:123. https://doi.org/10.1186/s13104-019-4159-6. factors among pregnant mother in Ambo District, West Shoa, Oromia, 32. EvansCEL,HutchinsonJ,ChristianMS,HancockN,CadeJE.Measuresoflow Ethiopia, 2018. Ethiop J Reprod Health. 2018;10(4):43–51. food variety and poor dietary quality in a crosssectional study of London school children. Eur J Clin Nutr. 2018. https://doi.org/10.1038/s41430-017-0070-1. 10. Gemeda D, Fekadu B, Wondu G, Habitamu F. Assessment of nutritional 33. Girard AW, Oludea O. Nutrition education and counselling provided during practice of pregnant mothers on maternal nutrition and associated factors pregnancy: effects on maternal, neonatal and child health outcomes. in Guta Gida Woreda, East Wollega Zone, Ethiopia. Star J. 2013;2(3):105-13. Paediatr Perinat Epidemiol. 2012;26(1):191–204. 11. Alemayehu MS, Tesema EM. Dietary practice and associated factors among 34. Thangaratinam S, Rogozinska E, Jolly K, Glinkowski S, Roseboom T, Tomlison pregnant women in Gondar town north west, Ethiopia. Int J Nutr Food Sci. JW, et al. Effect of interventions in pregnancy on maternal weight and 2015;4(6):707–12. https://doi.org/10.11648/j.ijnfs.20150406.27. obstetric outcomes: meta-analysis of randomised evidence. BMJ. 2012;344:1- 12. Nana A, Zema T. Dietary practices and associated factors during pregnancy 15:e2088. https://doi.org/10.1136/bmj.e2088. in northwestern Ethiopia. BMC Pregnancy Childbirth. 2018;18:183. 35. NchangMugyia AS, Tanya ANK, Njotang PN, Ndombo PK. Knowledge and 13. Tenaw Z, Arega M, Tachbele E. Nutritional knowledge, attitude and attitudes of pregnant mothers towards maternal dietary practices during practices among pregnant women who attend antenatal care at public pregnancy at the Etoug-Ebe Baptist Hospital Yaounde. Health Sci Dis. 2016; hospitals of Addis Ababa, Ethiopia. Int J Nurs Midwifery. 2018;10(7):81–9. 17(2):24–9. 14. Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, Md O, et al. Maternal 36. Kuche D, Singh P, Moges D. Dietary practice and associated factors among and child undernutrition and overweight in low-income and middle- pregnant women in Wondo Genet District, Southern Ethiopia. J Pharm Sci income countries. Lancet. 2013;382:427–51. Innov. 2015;4(5):270–5. 15. Abu-Saad K, Fraser D. Maternal nutrition and birth outcomes. Epidemiol Rev. 37. Luo B, Liu Y, Peng W, Feng X. Dietary behavior and influencing factors 2010;32(1):5–25. among rural pregnant women in poverty-stricken West China, Sichuan 16. Central Statistical Agency (CSA) [Ethiopia] and ICF. Ethiopia demographic Province: a cross-sectional study. Front Womens Heal. 2016;1(3):1–7. https:// and health survey 2016. Addis Ababa and Rockville: CSA and ICF; 2016. doi.org/10.15761/FWH.1000118. 17. Shin D, Bianchi L, Chung H, Weatherspoon L, Song WO. Is gestational 38. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. weight gain associated with diet quality during pregnancy? Matern Child – – 1991;50(2):179 211. https://doi.org/10.1016/0749-5978(91)90020-T. Health J. 2014;18:1433 43. 39. Nunnery DL, Labban JD, Dharod JM. Interrelationship between food security 18. Kubota K, Itoh H, Tasaka M, Naito H, Fukuoka Y, Kato KM, et al. Changes of status, home availability of variety of fruits and vegetables and their dietary maternal dietary intake, bodyweight and fetal growth throughout intake among low-income pregnant women. Public Health Nutr. 2018;21(4): pregnancy in pregnant Japanese women. J Obstet Gynaecol Res. 2013;39(9): 807–15. https://doi.org/10.1017/S1368980017003032. – 1383 90. 40. Na M, Sucheta M, Parul C, Hasmot A, Saijuddin S, Abu Ahmed S, et al. 19. Li N, Liu E, Guo J, Pan L, Li B, Wang P, et al. Maternal Prepregnancy body Maternal dietary diversity decreases with household food insecurity in rural mass index and gestational weight gain on pregnancy outcomes. PLoS Bangladesh: a longitudinal analysis. J Nutr. 2016;146:2109–16. One. 2013;8(12):e82310. https://doi.org/10.1371/journal.pone.0082310. 41. Anderman TL, Remans R, Wood SA, DeRosa K, DeFries RS. Synergies and 20. Federal Democratic Republic of Ethiopia. National Nutrition Program Multi- tradeoffs between cash crop production and food security: a case study in sectoral Implementation Guide. Addis Ababa: 2016. rural Ghana. Food Sec. 2014. https://doi.org/10.1007/s12571-014-0360-6. 21. Lindsay KL, Gibney ER, McAuliffe FM. Maternal nutrition among women from sub-Saharan Africa, with a focus on Nigeria, and potential implications for pregnancy outcomes among immigrant populations in developed Publisher’sNote countries. J Hum Nutr Diet. 2012;25:534–46. https://doi.org/10.1111/j.365- Springer Nature remains neutral with regard to jurisdictional claims in 277X.2012.01253.x. published maps and institutional affiliations.